User Guide

A step-by-step guide for using the AI-powered resume screening system built on Sentence-BERT (SBERT) and advanced NLP techniques.

Welcome to the Intelligent Resume Screening and Shortlisting System. This guide helps recruiters, administrators, and students understand how to use our platform effectively for candidate shortlisting using AI and NLP.

1. Overview

The system is designed to automatically analyze, rank, and shortlist resumes based on their semantic similarity to job descriptions using SBERT (Sentence-BERT) embeddings. Unlike traditional keyword searches, SBERT understands the context and meaning of words, improving fairness and efficiency in hiring.

2. System Features

3. Getting Started

  1. Step 1: Sign up or log in to your recruiter or admin account.
  2. Step 2: Navigate to the Job Description Upload section and upload the role details (title, skills, experience level, etc.).
  3. Step 3: Upload multiple candidate resumes (PDF or DOCX format).
  4. Step 4: Click “Analyze & Shortlist” to process the data through the NLP pipeline.
  5. Step 5: View the generated similarity scores and download the top-matched resumes.

4. How the System Works

The backend uses Sentence-BERT to encode resumes and job descriptions into numerical embeddings. These embeddings are then compared using cosine similarity to determine semantic closeness. The workflow includes:

5. Tips for Recruiters

6. Administrator Functions

7. Security & Data Handling

All data is encrypted using AES standards and stored in secure cloud environments. The system automatically deletes temporary resume data after analysis to maintain confidentiality. Admins can configure retention periods as per institutional policy.

8. Troubleshooting

9. Ethical Use Guidelines

This AI system is designed to assist recruitment, not replace human judgment. Always validate model outputs and avoid discriminatory decisions based solely on algorithmic results. The platform follows fairness and transparency principles aligned with responsible AI practices.

10. Support and Feedback

For assistance, contact the technical team via email or chat support. User feedback helps us enhance the SBERT model and improve interface usability.

Email: support@example.com
Feedback Portal: www.example.com/feedback